Google is planning to incorporate the Content Authenticity Initiative’s (CAI) C2PA provenance technology into its Search and Ads systems to better identify if an image is straight out of camera, edited, or created by AI.
C2PA provenance technology is metadata that is attached to an image and gives viewers the ability to tell if it was taken by a camera, edited by software, or produced with generative AI. The goal of the CAI and its C2PA content credentials is to help people stay more informed about the origins of an image. Google joined the CAI earlier this year and is a steering committee member along with Adobe, BBC, Intel, Microsoft, Publicis Groupe, Sony, and Truepic.
It is now planning to add Content Credentials to two of its products: Search and Ads. In Search, if an image contains C2PA metadata, users will be able to click on the “About this image” option — which is in Lens and Circle to Search, too — to see if it was created with or edited with AI tools. The Ads system will also add C2PA integration and Google says its “goal is to ramp this up over time and use C2PA signals to inform how we enforce key policies.”
One key point is that for this system to work, C2PA metadata must be present in the images in question. If there is no metadata, it’s not possible to see an image’s provenance. This has always been an issue with the rollout of C2PA, but the eventual hope is that users will try and only trust images that bear the C2PA metadata. At this time, however, the number of images that don’t have the metadata vastly outnumbers those that do.
That’s why Google says it is encouraging more hardware providers to adopt C2PA Content Credentials since the loop only works when everyone agrees to use the standard. Most camera manufacturers have signed on to the C2PA standard and are members of the CAI, but to date, very few cameras actively use the metadata — Leica’s M11-P is one example.
Google says it is “exploring” ways to give YouTube viewers visibility to C2PA information, too.
“We’re also exploring ways to relay C2PA information to viewers on YouTube when content is captured with a camera, and we’ll have more updates on that later in the year. We will ensure that our implementations validate content against the forthcoming C2PA Trust list, which allows platforms to confirm the content’s origin. For example, if the data shows an image was taken by a specific camera model, the trust list helps validate that this piece of information is accurate,” Google says.
“These are just a few of the ways we’re thinking about implementing content provenance technology today, and we’ll continue to bring it to more products over time.”
Google isn’t just betting on C2PA. It is also using SynthID and has joined a set of other coalitions and groups focused on AI safety and research such as the Secure AI Framework. That said, its application of AI in the new Pixel 9 series phones do not do a good job of highlighting the use of AI, which is especially concerning given how realistic its “Reimagine” edits are.
Image credits: Header photo licensed via Depositphotos.